Int. J. Appl. Math. Comput. Sci., 2003, Vol. 13, No.

3, 297–305

DIFFERENT MODELS OF CHEMOTHERAPY TAKING INTO ACCOUNT

DRUG RESISTANCE STEMMING FROM GENE AMPLIFICATION

JAROSŁAW ŚMIEJA∗ , A NDRZEJ ŚWIERNIAK∗

∗
Department of Automatic Control
Silesian University of Technology
ul. Akademicka 16, 44–100 Gliwice, Poland
e-mail: {jsmieja, aswierniak}@ia.polsl.gliwice.pl

This paper presents an analysis of some class of bilinear systems that can be applied to biomedical modelling. It com-
bines models that have been studied separately so far, taking into account both the phenomenon of gene amplification and
multidrug chemotherapy in their different aspects. The mathematical description is given by an infinite dimensional state
equation with a system matrix whose form allows decomposing the model into two interacting subsystems. While the first
one, of a finite dimension, can have any form, the other is infinite dimensional and tridiagonal. A methodology of the anal-
ysis of such models, based on system decomposition, is presented. An optimal control problem is defined in the l1 space.
In order to derive necessary conditions for optimal control, the model description is transformed into an integro-differential
form. Finally, biomedical implications of the obtained results are discussed.

Keywords: biomedical modelling, infinite dimensional systems, multivariable control

1. Introduction al., 1997a; 1998), studies of infinite dimensional models

Despite the long history of research and the rich litera-
ture devoted to modelling and control problems of infinite
dimensional systems, almost all efficient methods devel-
oped to deal with them present approaches suitable for
PDE models, and optimisation techniques are often lim-
ited to LQ problems. More general solutions, involving
abstract differential equations (Curtain and Zwart, 1995),
lead, in turn, to theoretical results whose applicability is
arguable.
Models based on an infinite number of state equa-
tions may be applied to a variety of systems. Besides the
models of drug resistance evolution caused by gene am-
plification (Kimmel and Axelrod, 1990; Polański et al.,
1997) analysed in this paper, they may also describe, e.g.,
RC ladders, which are approximations of long transmis-
sion lines (Zadeh and Desoer, 1963), microsatellite re-
peats evolution (Świerniak et al., 2002), which plays an
important role in genetic disorders (Ramel, 1997), telom-
ere shortening (Arino et al., 1995; Olofsson and Kimmel,
1999) responsible for cell aging and death, or some queu-
ing systems (Kleinrock, 1976). Usually, additional as-
sumptions are made, resulting in tridiagonal system ma-
trices. Moreover, the analysis of such models is often lim-
ited to their finite-dimensional approximations. However,
in that case, some dynamical properties may be neglected.
Moreover, as shown in our previous papers (Świerniak et
may lead to compact results, convenient in further analy-
sis, which would be impossible or very difficult to obtain
in a finite dimensional approximation.
Our previous works (e.g., Świerniak et al., 1999;
Śmieja et al., 1999) dealt with models with tridiago-
nal system matrices. They led to the development of a
methodology for investigating such systems and formed a
basis for further generalisation. This work pushes the re-
search a step further, studying the properties of a model, in
which significantly fewer simplifications have been made
and fewer additional assumptions are required. Moreover,
it combines models that have been studied separately so
far, taking into account the phenomenon of both gene am-
plification and multidrug chemotherapy in their different
aspects.
Three different examples are discussed in this paper,
each of them addressing different aspects of cancer cell
modelling. As the first one, a model taking into account
the partial sensitivity of the resistant subpopulation will
be introduced. In this case, it is assumed that the resis-
tant subpopulation consists of two parts: the one which is
sensitive to the drug (but, unlike in previous works, may
contain cells of a different drug sensitivity), and the other,
which is completely drug-resistant.
Subsequently, an attempt to model multidrug proto-
cols will be presented. The motivation behind it is that
 298 J. Śmieja and A. Świerniak

most of the existing forms of therapy consist in using
several drugs instead of a single one. Then, modelling
should take into account increasing drug resistance to each
chemotherapeutic agent used.
Finally, the phase-specific control of the drug-
sensitive cancer population will be addressed. Actually,
each drug affects cells being in particular phases and it
makes sense to combine these drugs so that their cu-
mulative effect on the cancer population is the greatest.
So far, phase-specific chemotherapy has been considered
without any regard to problems stemming from increas-
ing drug resistance. Combining the infinite dimensional
model of drug resistance with the phase-specific model
of chemotherapy should bring mathematical modelling
much closer to its clinical application.
populations is presented in Fig. 1(a).
2. A cell of Type i for i ≥ 1 may mutate in a short
time interval (t, t + dt) into a Type i + 1 cell with
probability bi dt + o(dt) and into a Type i − 1 cell
with probability di dt + o(dt). A cell of Type i = 0
may mutate in a short time interval (t, t + dt) into
a Type 1 cell with probability α dt + o(dt), where
α is several orders of magnitude smaller than any of
bi ’s and di ’s.
3. The drug action results in a fraction ui of ineffective
divisions in the cells of Type i (hence 0 ≤ ui ≤ 1).
4. The process is initiated at time t = 0 by a finite
population of cells of different types.
A graph representing the possible flows between sub-

2. Original Mathematical Model

0 0 α02
d10
The original model and its properties were thoroughly dis- d1 α α01 d20
cussed, e.g., in (Świerniak et al., 1998; 1999). However, 0 1
1
1 d30 α03 2 1–γ
the underlying biological background remains the same
α13 d32
also for the subject of this paper and therefore it needs to d2 b1 α23 γ
d31 d
be briefly introduced. 3 2
2
In this section some model of a cell population with
d3 b2 d b d b
evolving drug resistance caused by gene amplification or
... ...
other mechanisms is presented. Based on some results ...
of (Kimmel and Axelrod, 1990; Axelrod et al., 1994; di bi-1 d b d b
Harnevo and Agur, 1993), the model is general enough i i i
to allow various interpretations. bi b b
di+1 d d
We consider a population of neoplastic cells stratified ... ... ...
into subpopulations of cells of different types, labelled
with the numbers i = 0, 1, 2, . . . . If the biological process (a) (b) (c)
considered is gene amplification, then the cells of differ-
ent types are identified with different numbers of copies of Fig. 1. Flows between subpopulations for the model taking into
the drug resistance gene and varying levels of resistance. account (a) the original assumptions or partial sensitivity
Cells of Type 0, with no copies of the gene, are sensi- of the resistant subpopulation, (b) two-drug chemother-
apy, and (c) phase-specific chemotherapy; in all the
tive to the cytostatic agent. Due to the mutational event,
cases the numbers denote cell types.
a sensitive cell of Type 0 can acquire a copy of the gene
that makes it resistant to the agent. Likewise, the division
of resistant cells can result in a change in the number of If we denote by Ni (t) the expected number of cells
gene copies. The resistant subpopulation consists of cells of Type i at time t, and we assume the simplest case, in
of Types i for 1, 2, . . . . The probability of the mutational which the resistant cells are insensitive to the drug’s ac-
event in a sensitive cell is several orders smaller than the tion, and there are no differences between the parameters
probability of the change in the number of gene copies in of cells of different types, the model is described by the
a resistant cell. Since we do not limit the number of gene following system of ODE’s:
copies per cell, the number of different cell types is denu-   
merably infinite. 
 Ṅ0 (t) = 1 − 2u0 (t) λ0 N0 (t) − αN0 (t) + d1 N1 (t),

  
Cell division and the change in the number of gene



 Ṅ 1 (t) = 1 − 2u 1 (t) λ1 N1 (t) − (b1 + d1 )N1 (t),
copies are stochastic processes with the following hy-


 +d2 N2 (t) + αN0 (t),
potheses: .. (1)

 .
1. The lifespans of all cells are independent exponen-


  
tially distributed random variables with means 1/λl



 Ṅi (t) = 1 − 2ui (t) λi Ni (t) − (bi + di )Ni (t)

for cells of Type i. +di+1 Ni+1 (t) + bi−1 Ni−1 (t) for i ≥ 2.

Different models of chemotherapy taking into account drug resistance stemming from gene amplification 299

So far, only the simplest case has been investigated, realistic, it is sensible to find the optimal control which
in which the resistant cells are completely insensitive to minimises the performance index
the drug’s action and there are no differences between the
l−1 ∞ m Z T
parameters of cells of different types: X X X
J = Ni (T )+r1 Ni (T ) + r uk (τ ) dτ , (4)
   i=0 i=l k=0 0

 Ṅ0 (t) = 1 − 2u(t) λN0 (t) − αN0 (t) + dN1 (t),

 where r1 and r are non-negative weighting factors.


 Ṅ1 (t) = λN1 (t) − (b + d)N1 (t) + dN2 (t)

 The idea on which such optimisation is based is to

 +αN0 (t),
(2) minimise the resistant cancer subpopulation at the end of
 .
.. the therapy while minimizing the negative cumulative ef-



 fect of the drug represented by the integral component.



 Ṅi (t) = λNi (t) − (b + d)Ni (t) + dNi+1 (t)

+bNi−1 (t) for i ≥ 2.

3. Model Taking into Account the Partial
However, using the same line of reasoning that has Sensitivity of the Resistant Subpopulation
been applied to that case, it is also possible to analyse a In this case, it is assumed that the resistant subpopulation
less simplified model. If it is assumed that the parameters consists of two parts: the one, which is partially sensi-
may vary for a given finite number of cells and are the tive to the drug, and the other, which is completely drug-
same only for the infinite dimensional tail of the system, resistant. Then the following set of equations is obtained:
the following model can be investigated:
  
   
 Ṅ0 (t) = 1−2u(t) λ0 N0 (t)−αN0 (t)+d1 N1 (t),
 Ṅ0 (t) = 1 − 2u0 (t) λ0 N0 (t) − αN0 (t) 

  
Ṅ1 (t) = 1−2µ1 u(t) λ1 N1 (t)−(b1 +d1 )N1 (t)

 



 +d1 N1 (t), 

+d2 N2 (t) + α N0 (t),

 

   
Ṅ1 (t) = 1 − 2u1 (t) λ1 N1 (t) − (b1 + d1 )N1 (t)
 
.
 
..

 

 
+d2 N2 (t) + αN0 (t),

 

   
Ṅl−1 (t) = 1 − 2µl−1 u(t) λl−1 Nl−1 (t)

..




 . (5)
 
 −(bl−1 + dl−1 )Nl−1 (t) + dl Nl (t)
(3)
 
Ṅl−1 (t) = 1 − 2ul−1 (t) λl−1 Nl−1 (t)


+bl−2 Nl−2 (t),
 


−(bl−1 + dl−1 )Nl−1 (t) + dl Nl (t)
 
..

 

 
.
 
+bl−2 Nl−2 (t),

 


 
Ṅi (t) = λNi (t) − (b + d)Ni (t) + dNi+1 (t)

..

 


.

 

+bNi−1 (t) for i ≥ l,

 
Ṅi (t) = λNi (t) − (b + d)Ni (t) + dNi+1 (t)






+bNi−1 (t) for i ≥ l. where the µi ’s satysfying 0 ≤ µi ≤ 1 are “efficiency fac-
tors”, determining the effectiveness of the drug in relation
to a particular type of cell. Due to the general assumptions
Moreover, multivariable control is allowed, meaning about the model, which were presented at the beginning of
that either some types of the resistant cells can be affected this section, these factors satisfy the following relations:
by chemotherapy or different drugs are being used. A jus-
tification of its usefulness is presented in the following 0 ≤ µi ≤ µi−1 ≤ 1, i = 1, 2, . . . , l − 1. (6)
sections.
Several control problems arising in all these cases
can be addressed based on the model. One of them is 4. Multidrug Protocols
establishing constant control u (in that case it leads to
the determination of feedback parameters) that stabilises Most of the existing forms of therapies consist in using
the infinite dimensional system. In biological terms, it several drugs instead of a single one. Coldman and Goldie
refers to calculating a constant dose of the chemothera- (1986) suggested that a such chemotherapy might reduce
peutic agent that suppresses the growth of the resistant the drug resistance effects. Then the modelling should
subpopulation. However, the constant treatment protocol, take into account the increasing drug resistance to each
which guarantees the decay of the cancer population af- of the chemotherapeutic agents used. There is no known
ter a sufficiently long time, is not realistic. First of all, it mathematical approach to analysing such a problem.
does not take into account the cumulated negative effect of Let us consider the case of a simultaneous use of two
the drug upon normal tissues. To make the solution more types of drugs. Assume the simplest case, in which the
 300 J. Śmieja and A. Świerniak

resistance of the cells means that they are insensitive to
the drugs’ actions, and there are no differences between
the parameters of cells of different types. Then we could
distinguish four different subpopulations of cells: Type 0,
which is sensitive to both drugs, Type 1 and Type 2, sen-
sitive only to the first and the second agent, respectively,
and Type i ≥ 3, resistant to both drugs. The second hy-
pothesis for the original model has to be modified in the
following way:
• A cell of Type 0 may mutate in a short time inter-
val (t, t + dt) into a Type 1 cell with probabilities
α13 dt + o(dt) or into a Type 2 cell with probability
α02 dt + o(dt) or into a Type 3 cell with probability
α03 dt + o(dt).
• Each cell of Type 1 or Type 2 may mutate in a short
time interval (t, t + dt) into a Type 3 cell with
probabilities α13 dt + o(dt) and α23 dt + o(dt), re-
spectively, or into a Type 0 cell with probabilities
d10 dt + o(dt) and d20 dt + o(dt), respectively.
• A cell of Type 3 may mutate in a short time interval
(t, t + dt) into a Type 0, Type 1 or a Type 2 cell
with probabilities d30 dt+o(dt), d31 dt+o(dt) and
d32 dt+o(dt), respectively, or into a Type 4 cell with
probability b dt + o(dt).
Although in the mathematical model given above it is
assumed that the mechanism for resistance is independent
for each drug, the methodology developed makes it pos-
sible to analyse also the case when different drugs affect
the same gene simultaneously. In that case, the model (7)
should be combined with (5). In fact, this combination is
possible in the general model introduced in Section 6.
5. Phase-Specific Control of the Drug-
Sensitive Cancer Population
The cell cycle is composed of a sequence of phases un-
dergone by each cell from its birth to division. Actu-
ally, each drug affects the cell being in a particular phase
and it makes sense to combine these drugs so that their
cumulative effect on the cancer population is the great-
est. So far, phase-specific chemotherapy has been con-
sidered only in the finite-dimensional case, without any
regard to problems stemming from increasing drug resis-
tance (Świerniak et al., 1997a; Swan, 1990). Combining
the infinite dimensional model of drug resistance with the
phase-specific model of chemotherapy should bring math-
ematical modelling much closer to its clinical application.

• A cell of Type i, i ≥ 4, may mutate in a short time Once again, some modification of the assumptions
interval (t, t + dt) into a Type i + 1 cell with prob- underlying the mathematical model presented at the be-
ability b dt + o(dt) and into Type i − 1 cell with ginning of this section should be introduced. The sensitive
probability d dt + o(dt), where αlϕ is several or- subpopulation consists of two types of cells: Type 0, being
ders of magnitude smaller than b and d. in the phase G1 +S, and Type 1, being in the phase G2 M .
The phase-specific drug affects only cells of Type 1. Then
A graph illustrating the possible transfers between the following set of equations can represent the system
different subpopulations is presented in Fig. 1(b). The fol- dynamics:
lowing set of equations is obtained:

Ṅ0 (t) = −λ0 N0 (t) + 2(1 − γ)[1 − u(t)]λ1 N1 (t)
  

 Ṅ0 (t) = 1 − β0 u0 (t) − β1 u1 (t) λN0 (t) 


+dN2 (t),
 
−(α01 + α02 + α02 )N0 (t) + d10 N1 (t)

 


 

 
+d20 N2 (t),  Ṅ1 (t) = −λ1 N1 (t) + λ 0 N0 (t),

 


 

   

Ṅ1 (t) = 1 − 2u0 (t) λN1 (t) − (α13 + d10 )N1 (t) Ṅ2 (t) = λ2 N2 (t) − (b + d)N1 (t)

 
(8)


  
+d31 N3 (t) + α01 N0 (t), +2γ 1 − u(t) λ1 N1 (t) + bN3 (t),

 

 

..

   

 Ṅ2 (t) = 1 − 2u1 (t) λN2 (t)
 

 


 .
−(α23 + d20 )N2 (t) + d32 N3 (t) (7) λNi (t) − (b + d)Ni (t) + dNi+1 (t)



 Ṅ i (t) =
 
+α02 N0 (t), +bNi−1 (t) for i ≥ 3,

 




Ṅ3 (t) = λN3 (t) − (b + d30 + d31 + d32 )N3 (t)




where γ is the probability of the primary mutational

+dN4 (t) + α13 N1 (t) + α23 N2 (t),



event. A graph illustrating possible transfers between dif-



 .
..
ferent subpopulations is presented in Fig. 1(c). Similarly,




Ṅi (t) = λNi (t) − (b + d)Ni (t) + dNi+1 (t) a multidrug therapy including blocking drugs (Świerniak




et al., 1997b; Brown and Thompson, 1975) as well as the

+bNi−1 (t) for i ≥ 4,

killing agent could be analysed in the same way, as pre-
where β1 and β2 are efficiency factors, and β0 +β1 ≤ 2. sented in the subsequent sections.
Different models of chemotherapy taking into account drug resistance stemming from gene amplification 301

6. General Mathematical Model 7. Model Decomposition

To make an analysis of the model, it is convenient to
The system is described by the following state equation: present it in the form of a block diagram shown in Fig. 2,
! effectively decomposing the model into two parts. The
m
X first one, of the finite dimension, does not require that
Ṅ = A+ ui B i N, (9) parameters meet any particular assumptions. The second
i=0
subsystem is infinite dimensional, with a tridiagonal sys-
tem matrix, and does not include terms containing control
where N = [N0 N1 N2 . . . Ni . . . ]T is an infinite variables ui (t).
dimensional state vector,
l −1 m l −1
.. N0 = ∑∑b j
0 ,i u i (t ) N j (t ) + ∑a 0, i N i (t )
 
à . 0 j =0 i =0 i =0
A =  . . 1. . . .. . . . .1.  , (10)

0 2 .. Ã2 l −1 m l −1

N l (t )
N l −2 = ∑∑b j
0 ,i u i (t ) N j (t ) + ∑a l − 2 , i N i (t ) N l −1 (t )
j =0 i =0 i =0
l −1 m l
. ∑∑b j
∑a
 
N l −1 = 0 ,i u i (t ) N j (t ) + l −1, i N i (t )
B̃ i .. 0 1 j =0 i =0 i =0
B =  . . . . . .· . . . . .  , (11)
03

 
a00 a01 ··· a0,l−1 0

 a10 a11 ··· a1,l−1 0

 N l (t ) = c1 N l −1 (t ) + a2 N l (t ) + a3 Nl +1 (t )
Ã1 =  .. .. .. , N l +1 (t ) = a1 N l (t ) + a2 N l +1 (t ) + a3 N l + 2 (t )
···

0

 . . . 

al−1,0 al−1,1 · · · al−1,l−1 al−1,l N i (t ) = a1 N i −1 (t ) + a2 N i (t ) + a3 N i +1 (t )

 
c1 a2 a3 0 0 0 ··· Fig. 2. Decomposition of the system model.
0 a1 a2 a3 0 0 ···
 
 
Ã2 =  ,

 0 0 a1 a2 a3 0 ··· 

.. .. .. .. .. .. .. 7.1. Infinite Dimensional Subsystem
. . . . . . .
First, consider the infinite dimensional tail without the in-
  fluence of cells Nl−1 :
bi0,0 bi0,1 ... bi0,l−1 
 i
 b1,0 bi1,1 ... bi1,l−1

 
 Ṅl (t) = a2 Nl (t) + a3 Nl+1 (t),
B̃ i =  , 
.. .. ..

Ṅ (t) = a1 Nl (t) + a2 Nl+1 (t) + a3 Nl+2 (t),


 l+1

...

 . . .  

i
..
bl−1,0 bil−1,1 ... bil−1,l−1 . (12)


 Ṅi (t) = a1 Ni−1 (t) + a2 Ni (t) + a3 Ni+1 (t),


..

u(t) stands for the m-dimensional control vector u =


.

[u0 u1 u2 . . . um−1 ]T , 01 , 02 and 03 are zero matri-
ces of dimensions l × ∞, ∞ × (l + 1) and ∞ × ∞, Using methods similar to that outlined in our previ-
respectively, l > m. ous works devoted to biomedical modelling (Świerniak
It is important to note that the model parameters sat- et al., 1999; Śmieja et al., 2000), it is possible to show
isfy the relations a3 > a1 > 0 and a2 < 0. However, a that for the initial condition Ni (0) = δik (the Kronecker
full problem analysis can be made in other possible cases delta), i.e., Nk (0) = 1, Ni (0) = 0 for i 6= k, the follow-
(e.g., when no additional conditions are to be satisfied by ing relations hold true:
the parameters a1 and a3 ) using exactly the same line p !k−l+1
of reasoning. The performance index to be minimised is k 1 s−a2 − (s−a2 )2 −4a1 a3
Nl (s) = , (13)
given by (4). a3 2a1
 302 J. Śmieja and A. Świerniak

1 Let us now introduce the following notation:

NΣk (s) =
s − (a1 + a2 + a3 )  
 p !k−l+1 0
s−a2 − 2
(s−a2 ) −4a1 a3  .. 
× 1− , (14) 
B̂ 1 =  . 
, C = [0, . . . , 0, 1] . (21)
2a1 
0
 | {z }
 
l components
where Nlk (s) and NP k
Σ (s) constitute the Laplace trans-
al−1,l
forms of Nl (t) and i≥1 Nik (t) = NΣk (t), respectively
k

(the superscript k is introduced to emphasize the index of Then, applying the standard control theory techniques
the state variable with a non-zero initial condition). Now, (Zadeh and Desoer, 1963), the following relation holds
assume that k = l. Then, after calculating the inverse true for u(t) = 0:
Laplace transform, the following formulae are obtained:
√ Xl−1 (s)
1
r 
a3 I1 2 a1 a3 t

K2 (s) = = C(sI − Ã1 )−1 B̂ 1 . (22)
l
Nl (t) = exp(a2 t), (15) Xl (s)
a3 a1 t
X Taking into account the linear form of this system, it
NΣl (t) =
 
Ni (t) = exp (a1 + a2 + a3 )t is possible to present the model in the form of a block di-
i̇≥l agram, shown in Fig. 3. This makes it possible to analyse
 r  Z t √ the dynamical properties of the closed-loop system.
a3 I1 (2 a1 a2 τ )
× 1−
a1 τ
0
Nl (s) Nl−1(s)
K2(s)

 
× exp − (a1 + a3 )τ dτ , (16)

where I1 (t) denotes the modified Bessel function of the

first order.
K1(s)
It should be emphasised that the assumption about
the initial condition does not introduce any additional con-
straints to the model applicability. Due to the linearity Fig. 3. Block diagram of the system without control.
of the infinite dimensional tail, any finite non-zero initial
condition can be incorporated into the final solution.
Using an asymptotic expansion of (16) and assum- 7.3. Case of a Constant Control Vector
ing a3 ≥ a1 , it was found (Polański et al., 1997) that the
stability condition for the autonomous system is Let us now consider the problem of stabilizing the sys-
√ tem (9) by constant control. Then the transfer function
a2 ≤ −2 a1 a3 . (17)
K2 (s) representing the finite dimensional subsystem in
Fig. 3 takes the following form:
7.2. Analysis of the Complete Model without Control Xl−1 (s)
K2 (s) =
The relation (6) can be used to determine the following Xl (s)
transfer function in the model (9):   m
X −1
Nl (s) = C sI − Ã1 + B̃ i B̂ 1 . (23)
K1 (s) = i=0
Nl−1 (s)
p
c1 s − a2 − (s − a2 )2 − 4a1 a3 Again, the standard control theory techniques, in-
= . (18)
a3 2a1 cluding the Nyquist criterion (Zadeh and Desoer, 1963),
Moreover, can be applied to find stability conditions for such a sys-
X tem.
Ni (t) = NΣl (t) + N + (t), (19)
i≥l

where 8. Integro-Differential Model

Z t
N + (t) = c1 NΣl (t − τ )Nl−1 (τ ) dτ (20) The system description (9) in the form of an infinite num-
0 ber of ODEs is not very convenient, although it can be
and NΣl (t) is defined by (16). used in different approaches to optimisation problems that
Different models of chemotherapy taking into account drug resistance stemming from gene amplification 303

will be considered in the next section. Instead, a model The goal is to minimise the performance index given
transformation into an integro-differential one is proposed by (4). Due to the particular form of both the performance
in this section. index and the equation governing the model, it is possible
Set   to find the solution to the problem applying the appropri-
N0 ate version of Pontryagin’s maximum principle (Pontrya-
 .  gin et al., 1962).
x̃ =  .  (24)
 .  It is important to notice that, although the perfor-
Nl−1 mance index (4) seems to consist of two components
and C k = [cj ], ck = 1, cj = 0 for j 6= k, i = (a sum and an integral), the sum actually involves another
1, 2, . . . , l − 1. Let us also assume that Ni (0) = 0 for integral, which stems from (19) and (20). Therefore, it
i > l − 1 (once again it should be stressed that any finite should be rewritten to emphasise this relation:
non-zero initial condition can be incorporated into the fi- l−1
X
nal solution). Then the last equation in the first subsystem, J= Ni (T ) + r1 NΣl (T )
influenced directly by control, as presented in Fig. 2, can i=0
Z T m 
be transformed into the integro-differential form l
X
+ r1 c1 NΣ (T −τ )Nl−1 (τ )+r uk (τ ) dτ . (30)
l−1 X
m l−1 0 k=0
X X
Ṅl−1 (t) = bjl−1,i ui (t)Nj (t) + al−1,i Ni (t) A number of formulations of necessary conditions
j=0 i=0 i=0
for the optimisation problem for dynamical systems gov-
Z t erned by integro-differential equations can be found in the
+ al−1,l k1 (t − τ )Nl−1 (τ ) dτ , (25) literature, e.g., (Bate, 1969; Connor, 1972; Gabasov and
0 Kirilowa, 1971). However, they are usually too general to
where k1 (t) is the inverse Laplace transform of K1 (s) be efficiently applied to such a particular problem or they
given by (18). have too strong constraints (e.g., regarding the smooth-
ness of the control function). Nevertheless, following the
Similarly, other equations can also be rewritten in the
line of reasoning presented by Bate (1969), it is possible
same way, leading to the transformation of the model (9)
to derive the necessary conditions for optimal control:
into the following form:
 X m
opt
uk (t) + pT (t)h(u, x̃)
Z t
u (t) = arg min r
˙˜x = h(u, x̃) + f˜(x̃, t, τ ) dτ , x̃(0) = x̃0 , (26) u
k=0
0 Z T 
where h(·, ·) are f˜(·, ·, ·) are the respective l-dimension- +al−1,l pl−1 (τ )k1 (t−τ )Nl−1 (τ ) dτ , (31)
t
al vector functions: 
l−1 X
X m l−1
X ṗT (t) = − q T (t) + pT (t)hx̃ (u, x̃)
hk (u, x̃) = bjk,i ui (t)Nj (t) + ak,i Ni , (27)
j=0 i=0 i=0
Z T 
+ T ˜
p (τ )fx̃ (t − τ )dτ , (32)
( t
0 for k < l−1,
f˜k (x̃, t, τ ) = (28)  T
al−1,l k1 (t−τ )Nl−1 (t) for k = l−1. q(t) = 0 . . . 0 r1 c1 NΣl (T − t) ,

9. Necessary Conditions for Optimal Control
After the transformation of the system description pre-
sented in the previous section, it is possible to effectively
address the resulting optimal control problem.
Let the system be governed by Eqn. (9), which is then
transformed into the form (26). The control is bounded,
i.e.,
0 ≤ uk (t) ≤ 1, (29)
where uk (t) = 1 represents the maximum allowable dose
of the drug k and uk (t) = 0 represents no application of
the drug k.
pi (T ) = 1, i = 0, 1, . . . , l − 1, (33)
p(t) being the adjoint vector.
Taking into account the constraint (29) and the bi-
linear form of (27), it can be proved that, in order to sat-
isfy (31), optimal control must be of the bang-bang type.
Then, to find an optimal number of switches and switch-
ing times, a gradient method can be developed, following
the line of reasoning presented in (Śmieja et al., 1999).
10. Conclusions
This paper is concerned with an infinite dimensional bi-
linear model of dynamical systems. Based on model de-
 304
composition, it is possible to analyse analytically some of
their dynamical properties. The transformation of system
description into one integro-differential equation allows
solving an optimal control problem with the performance
index defined in the l1 space of summable sequences.
One of the possible model applications is the mod-
elling of drug resistance emergence in cancer cells and
the analysis of possible chemotherapy protocols. Until
now, treatment protocols have been designed mainly on
the basis of experimental results and the general knowl-
edge about the drug activity. However, no general math-
ematical approach exists which would help to explain the
obtained results or to design a treatment in chemotherapy.
The results of this work can be used to reveal the desired
form of an optimal treatment protocol and to give some
hints as to its development. It can be also employed in
a qualitative analysis of the chosen protocol, taking into
account both the increasing resistance to chemoterapeu-
tic agents and allowing multidrug and phase-specific treat-
ments.
The main application of the obtained results is in the
field of biomedical modelling, from which the form of the
performance index in the analysed problem stems. Never-
theless, taking into account the broad class of systems de-
scribed by the presented model, it should easily be adapted
for usage in other areas.
Acknowledgement
This work was partially supported by the grant of the
State Committee for Scientific Research (KBN) in Poland
No. 4T11F 01824.
References
Arino O., Kimmel M. and Webb G.F. (1995): Mathematical
modelling of the loss of telomere sequences. — J. Theor.
Biol., Vol. 177, No. 1, pp. 45–57.
Axelrod D.E., Baggerly K.A. and Kimmel M. (1994): Gene am-
plification by unequal chromatid exchange: Probabilis-
tic modelling and analysis of drug resistance data. — J.
Theor. Biol., Vol. 168, No. 2, pp. 151–159.
Bate R.R. (1969): The optimal control of systems with transport
lag, In: Advances in Control and Dynamic Systems (C.T.
Leondes, Ed.). — Academic Press, Vol. 7, pp. 165–224.
Brown B.W. and Thompson J.R. (1975): A rationale for syn-
chrony strategies in chemotherapy, In: Epidemiology
(D. Ludwig and K.L. Cooke, Eds.). — Philadelphia:
SIAM Publ., pp. 31–48.
Coldman A.J. and Goldie J.H. (1986): A stochastic model for the
origin and treatment of tumors containing drug-resistant
cells. — Bull. Math. Biol., Vol. 48, No. 3–4, pp. 279–292.
J. Śmieja and A. Świerniak
Connor M.A. (1972): Optimal control of systems represented by
differential-integral equations. — IEEE Trans. Automat.
Contr., Vol. AC-17, No. 1, pp. 164–166.
Curtain R.F. and Zwart H.J. (1995): An Introduction to Infinite-
Dimensional Linear Systems Theory. — New York:
Springer.
Gabasov R. and Kirilowa F.M. (1971): Qualitative theory of op-
timal processes. — Moscow: Nauka, (in Russian).
Harnevo L.E. and Agur Z. (1993): Use of mathematical models
for understanding the dynamics of gene amplification. —
Mutat. Res., Vol. 292, No. 1, pp. 17–24.
Kimmel M. and Axelrod D.E. (1990): Mathematical models of
gene amplification with applications to cellular drug re-
sistance and tumorigenicity. — Genetics, Vol. 125, No. 3,
pp. 633–644.
Kleinrock L. (1976): Queuing Systems. Vol. 1: Theory. — New
York: Wiley.
Olofsson P. and Kimmel M. (1999): Stochastic models of telom-
ere shortening. — Math. Biosci., Vol. 158, No. 1, pp. 75–
92.
Polański A., Kimmel M. and Świerniak A. (1997): Qualitative
analysis of the infinite-dimensional model of evolution of
drug resistance, In: Advances in Mathematical Population
Dynamics—Molecules, Cells and Man (O. Arino, D. Ax-
elrod and M. Kimmel, Eds.). — Singapore: World Scien-
tific, pp. 595–612.
Pontryagin L.S., Boltyanski V.G., Gamkrelidze R.V. and Mis-
chenko E.F. (1962): Mathematical Theory of Optimal Pro-
cesses. — New York: Wiley.
Ramel C. (1997): Mini- and microsatellites. — Environmental
Health Perspectives, Vol. 105, Suppl. 104, pp. 781–789.
Śmieja J., Duda Z. and Świerniak A. (1999): Optimal control
for the model of drug resistance resulting from gene am-
plification. — Prep. 14th IFAC World Congress, Beijing,
China, Vol. L, pp. 71–75.
Śmieja J., Świerniak A. and Duda Z. (2000): Gradient method
for finding optimal scheduling in infinite dimensional mod-
els of chemotherapy. — J. Theor. Med., Vol. 3, No. 1,
pp. 25–36.
Swan G.W. (1990): Role of optimal control theory in cancer
chemotherapy. — Math. Biosci., Vol. 101, No. 2, pp. 237–
284.
Świerniak A., Śmieja J., Rzeszowska-Wolny J. and Kimmel
M. (2001): Random branching walk models arising in
molecular biology—control theoretic approach. — Proc.
IASTED MIC Conf., Innsbruck, Austria, Vol. II, pp. 584–
589.
Świerniak A., Kimmel M., Polański A. and Śmieja J. (1997a):
Asymptotic properties of infinite dimensional model of
drug resistance evolution. — Proc. ECC’97, Brussels, TU-
A-C-4, CD-ROM.
Different models of chemotherapy taking into account drug resistance stemming from gene amplification 305

Świerniak A., Polański A., Duda Z. and Kimmel M. (1997b): Świerniak A., Polański A., Śmieja J., Kimmel M. and
Phase-specific chemotherapy of cancer: Optimisation of Rzeszowska-Wolny J. (2002): Control theoretic approach
scheduling and rationale for periodic protocols. — Biocy- to random branching walk models arising in molecular bi-
bern. Biomed. Eng., Vol. 16, No. 1–2, pp. 13–43. ology. — Proc. ACC Conf., Anchorage, pp. 3449–3453.

Świerniak A., Kimmel M. and Polański A. (1998): Infinite di- Zadeh L.A. and Desoer C.A. (1963): Linear System Theory. The
mensional model of evolution of drug resistance of can- State Space Approach. — New York: McGraw-Hill.
cer cells. — J. Math. Syst. Estim. Contr., Vol. 8, No. 1,
pp. 1–17.

Świerniak A., Polański A., Kimmel M., Bobrowski A. and

Śmieja J. (1999): Qualitative analysis of controlled drug
resistance model—inverse Laplace and semigroup ap-
proach. — Contr. Cybern., Vol. 28, No. 1, pp. 61–74.